signal --- Set handlers for asynchronous events

Source code: Lib/signal.py


This module provides mechanisms to use signal handlers in Python.

General rules

The signal.signal() function allows defining custom handlers to be executed when a signal is received. A small number of default handlers are installed: SIGPIPE is ignored (so write errors on pipes and sockets can be reported as ordinary Python exceptions) and SIGINT is translated into a KeyboardInterrupt exception if the parent process has not changed it.

A handler for a particular signal, once set, remains installed until it is explicitly reset (Python emulates the BSD style interface regardless of the underlying implementation), with the exception of the handler for SIGCHLD, which follows the underlying implementation.

On WebAssembly platforms wasm32-emscripten and wasm32-wasi, signals are emulated and therefore behave differently. Several functions and signals are not available on these platforms.

Execution of Python signal handlers

A Python signal handler does not get executed inside the low-level (C) signal handler. Instead, the low-level signal handler sets a flag which tells the virtual machine to execute the corresponding Python signal handler at a later point(for example at the next bytecode instruction). This has consequences:

  • It makes little sense to catch synchronous errors like SIGFPE or SIGSEGV that are caused by an invalid operation in C code. Python will return from the signal handler to the C code, which is likely to raise the same signal again, causing Python to apparently hang. From Python 3.3 onwards, you can use the faulthandler module to report on synchronous errors.

  • A long-running calculation implemented purely in C (such as regular expression matching on a large body of text) may run uninterrupted for an arbitrary amount of time, regardless of any signals received. The Python signal handlers will be called when the calculation finishes.

  • If the handler raises an exception, it will be raised "out of thin air" in the main thread. See the note below for a discussion.

Signals and threads

Python signal handlers are always executed in the main Python thread of the main interpreter, even if the signal was received in another thread. This means that signals can't be used as a means of inter-thread communication. You can use the synchronization primitives from the threading module instead.

Besides, only the main thread of the main interpreter is allowed to set a new signal handler.

Module contents

Modifié dans la version 3.5: signal (SIG*), handler (SIG_DFL, SIG_IGN) and sigmask (SIG_BLOCK, SIG_UNBLOCK, SIG_SETMASK) related constants listed below were turned into enums (Signals, Handlers and Sigmasks respectively). getsignal(), pthread_sigmask(), sigpending() and sigwait() functions return human-readable enums as Signals objects.

The signal module defines three enums:

class signal.Signals

enum.IntEnum collection of SIG* constants and the CTRL_* constants.

Nouveau dans la version 3.5.

class signal.Handlers

enum.IntEnum collection the constants SIG_DFL and SIG_IGN.

Nouveau dans la version 3.5.

class signal.Sigmasks

enum.IntEnum collection the constants SIG_BLOCK, SIG_UNBLOCK and SIG_SETMASK.

Disponibilité : Unix.

See the man page sigprocmask(2) and pthread_sigmask(3) for further information.

Nouveau dans la version 3.5.

The variables defined in the signal module are:

signal.SIG_DFL

This is one of two standard signal handling options; it will simply perform the default function for the signal. For example, on most systems the default action for SIGQUIT is to dump core and exit, while the default action for SIGCHLD is to simply ignore it.

signal.SIG_IGN

This is another standard signal handler, which will simply ignore the given signal.

signal.SIGABRT

Abort signal from abort(3).

signal.SIGALRM

Timer signal from alarm(2).

Disponibilité : Unix.

signal.SIGBREAK

Interrupt from keyboard (CTRL + BREAK).

Disponibilité : Windows.

signal.SIGBUS

Bus error (bad memory access).

Disponibilité : Unix.

signal.SIGCHLD

Child process stopped or terminated.

Disponibilité : Unix.

signal.SIGCLD

Alias to SIGCHLD.

Availability: not macOS.

signal.SIGCONT

Continue the process if it is currently stopped

Disponibilité : Unix.

signal.SIGFPE

Floating-point exception. For example, division by zero.

Voir aussi

ZeroDivisionError is raised when the second argument of a division or modulo operation is zero.

signal.SIGHUP

Hangup detected on controlling terminal or death of controlling process.

Disponibilité : Unix.

signal.SIGILL

Illegal instruction.

signal.SIGINT

Interrupt from keyboard (CTRL + C).

Default action is to raise KeyboardInterrupt.

signal.SIGKILL

Kill signal.

It cannot be caught, blocked, or ignored.

Disponibilité : Unix.

signal.SIGPIPE

Broken pipe: write to pipe with no readers.

Default action is to ignore the signal.

Disponibilité : Unix.

signal.SIGSEGV

Segmentation fault: invalid memory reference.

signal.SIGSTKFLT

Stack fault on coprocessor. The Linux kernel does not raise this signal: it can only be raised in user space.

Availability: Linux.

On architectures where the signal is available. See the man page signal(7) for further information.

Nouveau dans la version 3.11.

signal.SIGTERM

Termination signal.

signal.SIGUSR1

User-defined signal 1.

Disponibilité : Unix.

signal.SIGUSR2

User-defined signal 2.

Disponibilité : Unix.

signal.SIGWINCH

Window resize signal.

Disponibilité : Unix.

SIG*

All the signal numbers are defined symbolically. For example, the hangup signal is defined as signal.SIGHUP; the variable names are identical to the names used in C programs, as found in <signal.h>. The Unix man page for 'signal()' lists the existing signals (on some systems this is signal(2), on others the list is in signal(7)). Note that not all systems define the same set of signal names; only those names defined by the system are defined by this module.

signal.CTRL_C_EVENT

The signal corresponding to the Ctrl+C keystroke event. This signal can only be used with os.kill().

Disponibilité : Windows.

Nouveau dans la version 3.2.

signal.CTRL_BREAK_EVENT

The signal corresponding to the Ctrl+Break keystroke event. This signal can only be used with os.kill().

Disponibilité : Windows.

Nouveau dans la version 3.2.

signal.NSIG

One more than the number of the highest signal number. Use valid_signals() to get valid signal numbers.

signal.ITIMER_REAL

Decrements interval timer in real time, and delivers SIGALRM upon expiration.

signal.ITIMER_VIRTUAL

Decrements interval timer only when the process is executing, and delivers SIGVTALRM upon expiration.

signal.ITIMER_PROF

Decrements interval timer both when the process executes and when the system is executing on behalf of the process. Coupled with ITIMER_VIRTUAL, this timer is usually used to profile the time spent by the application in user and kernel space. SIGPROF is delivered upon expiration.

signal.SIG_BLOCK

A possible value for the how parameter to pthread_sigmask() indicating that signals are to be blocked.

Nouveau dans la version 3.3.

signal.SIG_UNBLOCK

A possible value for the how parameter to pthread_sigmask() indicating that signals are to be unblocked.

Nouveau dans la version 3.3.

signal.SIG_SETMASK

A possible value for the how parameter to pthread_sigmask() indicating that the signal mask is to be replaced.

Nouveau dans la version 3.3.

The signal module defines one exception:

exception signal.ItimerError

Raised to signal an error from the underlying setitimer() or getitimer() implementation. Expect this error if an invalid interval timer or a negative time is passed to setitimer(). This error is a subtype of OSError.

Nouveau dans la version 3.3: This error used to be a subtype of IOError, which is now an alias of OSError.

The signal module defines the following functions:

signal.alarm(time)

If time is non-zero, this function requests that a SIGALRM signal be sent to the process in time seconds. Any previously scheduled alarm is canceled (only one alarm can be scheduled at any time). The returned value is then the number of seconds before any previously set alarm was to have been delivered. If time is zero, no alarm is scheduled, and any scheduled alarm is canceled. If the return value is zero, no alarm is currently scheduled.

Disponibilité : Unix.

See the man page alarm(2) for further information.

signal.getsignal(signalnum)

Return the current signal handler for the signal signalnum. The returned value may be a callable Python object, or one of the special values signal.SIG_IGN, signal.SIG_DFL or None. Here, signal.SIG_IGN means that the signal was previously ignored, signal.SIG_DFL means that the default way of handling the signal was previously in use, and None means that the previous signal handler was not installed from Python.

signal.strsignal(signalnum)

Returns the description of signal signalnum, such as "Interrupt" for SIGINT. Returns None if signalnum has no description. Raises ValueError if signalnum is invalid.

Nouveau dans la version 3.8.

signal.valid_signals()

Return the set of valid signal numbers on this platform. This can be less than range(1, NSIG) if some signals are reserved by the system for internal use.

Nouveau dans la version 3.8.

signal.pause()

Cause the process to sleep until a signal is received; the appropriate handler will then be called. Returns nothing.

Disponibilité : Unix.

See the man page signal(2) for further information.

See also sigwait(), sigwaitinfo(), sigtimedwait() and sigpending().

signal.raise_signal(signum)

Sends a signal to the calling process. Returns nothing.

Nouveau dans la version 3.8.

signal.pidfd_send_signal(pidfd, sig, siginfo=None, flags=0)

Send signal sig to the process referred to by file descriptor pidfd. Python does not currently support the siginfo parameter; it must be None. The flags argument is provided for future extensions; no flag values are currently defined.

See the pidfd_send_signal(2) man page for more information.

Availability: Linux >= 5.1

Nouveau dans la version 3.9.

signal.pthread_kill(thread_id, signalnum)

Send the signal signalnum to the thread thread_id, another thread in the same process as the caller. The target thread can be executing any code (Python or not). However, if the target thread is executing the Python interpreter, the Python signal handlers will be executed by the main thread of the main interpreter. Therefore, the only point of sending a signal to a particular Python thread would be to force a running system call to fail with InterruptedError.

Use threading.get_ident() or the ident attribute of threading.Thread objects to get a suitable value for thread_id.

If signalnum is 0, then no signal is sent, but error checking is still performed; this can be used to check if the target thread is still running.

Raises an auditing event signal.pthread_kill with arguments thread_id, signalnum.

Disponibilité : Unix.

See the man page pthread_kill(3) for further information.

See also os.kill().

Nouveau dans la version 3.3.

signal.pthread_sigmask(how, mask)

Fetch and/or change the signal mask of the calling thread. The signal mask is the set of signals whose delivery is currently blocked for the caller. Return the old signal mask as a set of signals.

The behavior of the call is dependent on the value of how, as follows.

  • SIG_BLOCK: The set of blocked signals is the union of the current set and the mask argument.

  • SIG_UNBLOCK: The signals in mask are removed from the current set of blocked signals. It is permissible to attempt to unblock a signal which is not blocked.

  • SIG_SETMASK: The set of blocked signals is set to the mask argument.

mask is a set of signal numbers (e.g. {signal.SIGINT, signal.SIGTERM}). Use valid_signals() for a full mask including all signals.

For example, signal.pthread_sigmask(signal.SIG_BLOCK, []) reads the signal mask of the calling thread.

SIGKILL and SIGSTOP cannot be blocked.

Disponibilité : Unix.

See the man page sigprocmask(2) and pthread_sigmask(3) for further information.

See also pause(), sigpending() and sigwait().

Nouveau dans la version 3.3.

signal.setitimer(which, seconds, interval=0.0)

Sets given interval timer (one of signal.ITIMER_REAL, signal.ITIMER_VIRTUAL or signal.ITIMER_PROF) specified by which to fire after seconds (float is accepted, different from alarm()) and after that every interval seconds (if interval is non-zero). The interval timer specified by which can be cleared by setting seconds to zero.

When an interval timer fires, a signal is sent to the process. The signal sent is dependent on the timer being used; signal.ITIMER_REAL will deliver SIGALRM, signal.ITIMER_VIRTUAL sends SIGVTALRM, and signal.ITIMER_PROF will deliver SIGPROF.

The old values are returned as a tuple: (delay, interval).

Attempting to pass an invalid interval timer will cause an ItimerError.

Disponibilité : Unix.

signal.getitimer(which)

Returns current value of a given interval timer specified by which.

Disponibilité : Unix.

signal.set_wakeup_fd(fd, *, warn_on_full_buffer=True)

Set the wakeup file descriptor to fd. When a signal is received, the signal number is written as a single byte into the fd. This can be used by a library to wakeup a poll or select call, allowing the signal to be fully processed.

The old wakeup fd is returned (or -1 if file descriptor wakeup was not enabled). If fd is -1, file descriptor wakeup is disabled. If not -1, fd must be non-blocking. It is up to the library to remove any bytes from fd before calling poll or select again.

When threads are enabled, this function can only be called from the main thread of the main interpreter; attempting to call it from other threads will cause a ValueError exception to be raised.

There are two common ways to use this function. In both approaches, you use the fd to wake up when a signal arrives, but then they differ in how they determine which signal or signals have arrived.

In the first approach, we read the data out of the fd's buffer, and the byte values give you the signal numbers. This is simple, but in rare cases it can run into a problem: generally the fd will have a limited amount of buffer space, and if too many signals arrive too quickly, then the buffer may become full, and some signals may be lost. If you use this approach, then you should set warn_on_full_buffer=True, which will at least cause a warning to be printed to stderr when signals are lost.

In the second approach, we use the wakeup fd only for wakeups, and ignore the actual byte values. In this case, all we care about is whether the fd's buffer is empty or non-empty; a full buffer doesn't indicate a problem at all. If you use this approach, then you should set warn_on_full_buffer=False, so that your users are not confused by spurious warning messages.

Modifié dans la version 3.5: On Windows, the function now also supports socket handles.

Modifié dans la version 3.7: Added warn_on_full_buffer parameter.

signal.siginterrupt(signalnum, flag)

Change system call restart behaviour: if flag is False, system calls will be restarted when interrupted by signal signalnum, otherwise system calls will be interrupted. Returns nothing.

Disponibilité : Unix.

See the man page siginterrupt(3) for further information.

Note that installing a signal handler with signal() will reset the restart behaviour to interruptible by implicitly calling siginterrupt() with a true flag value for the given signal.

signal.signal(signalnum, handler)

Set the handler for signal signalnum to the function handler. handler can be a callable Python object taking two arguments (see below), or one of the special values signal.SIG_IGN or signal.SIG_DFL. The previous signal handler will be returned (see the description of getsignal() above). (See the Unix man page signal(2) for further information.)

When threads are enabled, this function can only be called from the main thread of the main interpreter; attempting to call it from other threads will cause a ValueError exception to be raised.

The handler is called with two arguments: the signal number and the current stack frame (None or a frame object; for a description of frame objects, see the description in the type hierarchy or see the attribute descriptions in the inspect module).

On Windows, signal() can only be called with SIGABRT, SIGFPE, SIGILL, SIGINT, SIGSEGV, SIGTERM, or SIGBREAK. A ValueError will be raised in any other case. Note that not all systems define the same set of signal names; an AttributeError will be raised if a signal name is not defined as SIG* module level constant.

signal.sigpending()

Examine the set of signals that are pending for delivery to the calling thread (i.e., the signals which have been raised while blocked). Return the set of the pending signals.

Disponibilité : Unix.

See the man page sigpending(2) for further information.

See also pause(), pthread_sigmask() and sigwait().

Nouveau dans la version 3.3.

signal.sigwait(sigset)

Suspend execution of the calling thread until the delivery of one of the signals specified in the signal set sigset. The function accepts the signal (removes it from the pending list of signals), and returns the signal number.

Disponibilité : Unix.

See the man page sigwait(3) for further information.

See also pause(), pthread_sigmask(), sigpending(), sigwaitinfo() and sigtimedwait().

Nouveau dans la version 3.3.

signal.sigwaitinfo(sigset)

Suspend execution of the calling thread until the delivery of one of the signals specified in the signal set sigset. The function accepts the signal and removes it from the pending list of signals. If one of the signals in sigset is already pending for the calling thread, the function will return immediately with information about that signal. The signal handler is not called for the delivered signal. The function raises an InterruptedError if it is interrupted by a signal that is not in sigset.

The return value is an object representing the data contained in the siginfo_t structure, namely: si_signo, si_code, si_errno, si_pid, si_uid, si_status, si_band.

Disponibilité : Unix.

See the man page sigwaitinfo(2) for further information.

See also pause(), sigwait() and sigtimedwait().

Nouveau dans la version 3.3.

Modifié dans la version 3.5: The function is now retried if interrupted by a signal not in sigset and the signal handler does not raise an exception (see PEP 475 for the rationale).

signal.sigtimedwait(sigset, timeout)

Like sigwaitinfo(), but takes an additional timeout argument specifying a timeout. If timeout is specified as 0, a poll is performed. Returns None if a timeout occurs.

Disponibilité : Unix.

See the man page sigtimedwait(2) for further information.

See also pause(), sigwait() and sigwaitinfo().

Nouveau dans la version 3.3.

Modifié dans la version 3.5: The function is now retried with the recomputed timeout if interrupted by a signal not in sigset and the signal handler does not raise an exception (see PEP 475 for the rationale).

Examples

Here is a minimal example program. It uses the alarm() function to limit the time spent waiting to open a file; this is useful if the file is for a serial device that may not be turned on, which would normally cause the os.open() to hang indefinitely. The solution is to set a 5-second alarm before opening the file; if the operation takes too long, the alarm signal will be sent, and the handler raises an exception.

import signal, os

def handler(signum, frame):
    signame = signal.Signals(signum).name
    print(f'Signal handler called with signal {signame} ({signum})')
    raise OSError("Couldn't open device!")

# Set the signal handler and a 5-second alarm
signal.signal(signal.SIGALRM, handler)
signal.alarm(5)

# This open() may hang indefinitely
fd = os.open('/dev/ttyS0', os.O_RDWR)

signal.alarm(0)          # Disable the alarm

Note on SIGPIPE

Piping output of your program to tools like head(1) will cause a SIGPIPE signal to be sent to your process when the receiver of its standard output closes early. This results in an exception like BrokenPipeError: [Errno 32] Broken pipe. To handle this case, wrap your entry point to catch this exception as follows:

import os
import sys

def main():
    try:
        # simulate large output (your code replaces this loop)
        for x in range(10000):
            print("y")
        # flush output here to force SIGPIPE to be triggered
        # while inside this try block.
        sys.stdout.flush()
    except BrokenPipeError:
        # Python flushes standard streams on exit; redirect remaining output
        # to devnull to avoid another BrokenPipeError at shutdown
        devnull = os.open(os.devnull, os.O_WRONLY)
        os.dup2(devnull, sys.stdout.fileno())
        sys.exit(1)  # Python exits with error code 1 on EPIPE

if __name__ == '__main__':
    main()

Do not set SIGPIPE's disposition to SIG_DFL in order to avoid BrokenPipeError. Doing that would cause your program to exit unexpectedly whenever any socket connection is interrupted while your program is still writing to it.

Note on Signal Handlers and Exceptions

If a signal handler raises an exception, the exception will be propagated to the main thread and may be raised after any bytecode instruction. Most notably, a KeyboardInterrupt may appear at any point during execution. Most Python code, including the standard library, cannot be made robust against this, and so a KeyboardInterrupt (or any other exception resulting from a signal handler) may on rare occasions put the program in an unexpected state.

To illustrate this issue, consider the following code:

class SpamContext:
    def __init__(self):
        self.lock = threading.Lock()

    def __enter__(self):
        # If KeyboardInterrupt occurs here, everything is fine
        self.lock.acquire()
        # If KeyboardInterrupt occurs here, __exit__ will not be called
        ...
        # KeyboardInterrupt could occur just before the function returns

    def __exit__(self, exc_type, exc_val, exc_tb):
        ...
        self.lock.release()

For many programs, especially those that merely want to exit on KeyboardInterrupt, this is not a problem, but applications that are complex or require high reliability should avoid raising exceptions from signal handlers. They should also avoid catching KeyboardInterrupt as a means of gracefully shutting down. Instead, they should install their own SIGINT handler. Below is an example of an HTTP server that avoids KeyboardInterrupt:

import signal
import socket
from selectors import DefaultSelector, EVENT_READ
from http.server import HTTPServer, SimpleHTTPRequestHandler

interrupt_read, interrupt_write = socket.socketpair()

def handler(signum, frame):
    print('Signal handler called with signal', signum)
    interrupt_write.send(b'\0')
signal.signal(signal.SIGINT, handler)

def serve_forever(httpd):
    sel = DefaultSelector()
    sel.register(interrupt_read, EVENT_READ)
    sel.register(httpd, EVENT_READ)

    while True:
        for key, _ in sel.select():
            if key.fileobj == interrupt_read:
                interrupt_read.recv(1)
                return
            if key.fileobj == httpd:
                httpd.handle_request()

print("Serving on port 8000")
httpd = HTTPServer(('', 8000), SimpleHTTPRequestHandler)
serve_forever(httpd)
print("Shutdown...")